Drafted with AI, reviewed by the 9Chain team.
Part 6 of 9 · 9Chain in Nine Parts
This part is for people who want to open a terminal. What sits under 9Chain, which tools you already know will work, where to point them, and how much failure the network is built to survive. Everything below runs on public test networks.
The layers, from your code down
The layers you build on are the same on both branches; only the engine at the bottom differs. That is what makes it safe to keep the engine question open: what you build above it is meant to outlive the decision underneath.
| Layer | What it holds |
|---|---|
| Applications | Solidity contracts, wallets, standard tooling |
| Sovereign chains | Your own validators, your own policy |
| Public surfaces | RPC, explorer, faucet |
| Shared platform | Chain creation, identity, operations tooling: the same on both branches |
| Engine (one of these becomes the mainnet) | 9Chain A1: avalanchego fork, Snow family consensus · 9Chain C1: Cosmos SDK, CometBFT consensus |
The order is alphabetical and carries no ranking. The community vote that decides which engine the official network uses is not open yet.
The documentation adds one boundary worth knowing (Platform architecture): the shared network is where identity, governance and value live for all chains, and its token is never a condition for a customer chain to run. A chain’s own fuel is kept separate from it.
A chain as a written description
On the C1 side, the docs describe a chain as a manifest. An operator reads it and keeps bringing the system back to that state: keys into a key store, a checked genesis, validators, RPC, faucet and explorer, then health checks on a cycle. Each loop gives the same result, so restarting the operator is always safe.
The hardest step is the genesis. As the docs put it, a wrong genesis cannot be fixed with a patch; it can only be fixed by creating the network again. A1 reaches the same outcome differently, with a chain factory on its P-Chain. Both were built to the same brief, which is what makes them comparable.
Standard EVM tooling
Both branches expose an EVM execution layer. Existing Solidity contracts, wallets and developer tools work without being rewritten for a virtual machine that exists only here. A new chain comes up with its own endpoint, and standard tooling connects to it the way it connects to any EVM chain.
Getting started on the test networks
Both branches publish the same set of public surfaces, so anyone can compare them. The values below are as published on 9chain.org (branch columns dated 2026-08-26):
| 9Chain A1 | 9Chain C1 | |
|---|---|---|
| Branch site | a1.9chain.org | c1.9chain.org |
| Explorer | a1.9scan.org | c1.9scan.org |
| JSON-RPC | rpc-a1.9chain.org/ext/bc/C/rpc | rpc-c1.9chain.org/jsonrpc |
| EVM chain ID | 9000000009 | 9000000000 |
| Faucet | Get test tokens — 9Chain Testnet A1 | 9Chain Faucet — get test LOVE9 in your browser |
| Node client | 9chaingo | love9d |
| Cross-chain messaging | Warp / ICM | IBC |
A shared explorer, 9scan.org, reads both networks in one place. The RPC addresses are addresses, not links: opening one in a browser tells you nothing. Point a wallet or a script at them instead.
Three things the docs ask you to keep in mind (Networks and access points):
- Take addresses from the branch site, then check the chain itself. Addresses can change, and one already has: an address with “testnet” in its name now serves a different chain from the one the name used to mean. Before trusting any number, ask the chain two questions: which network does it say it is, and when was its first block made? Those are things the chain states about itself and cannot rewrite.
- Both networks are run by the project itself, so every page describing them is the operator’s own account, not an independent review. The explorer is a separate project run by a different team: a second reader of the same ledger.
- These are test networks. Balances carry no value, chain state can be reset, and a network can be re-created from a new genesis without carrying balances over. The faucet hands out coins for the test, not assets. The endpoints are not promised to stay free or to handle any load. Treat what you build as disposable, and nothing here is an offer to buy or sell anything.
Security and fault tolerance, at a high level
From Security and fault tolerance:
One codebase, three operating modes: fully platform-run; hybrid (the default, which always includes a customer node and an auditor node run outside); or multi-party. They differ by one field in the manifest, not by code.
The fault rule is arithmetic. Consensus needs more than two thirds of voting power to finalise a block. With equal power, n nodes tolerate f sudden failures when n ≥ 3f+1. No configuration lowers that.
| Nodes | Failures tolerated |
|---|---|
| 3 | 0 |
| 4 | 1 |
| 5–6 | 1 |
| 7 | 2 (recommended for networks carrying real assets) |
| 10 | 3 |
Count machines, not validators. Nine validators on one cluster tolerate exactly one machine. No machine should hold a third or more of the signers, which means at least five machines that fail for different reasons.
Halt rather than split. Past the threshold, the chain stops instead of forking into two versions that both claim to be real. That is a safety choice, and it comes with a duty to recover quickly, with targets, procedures and drills.
What is not done yet, said plainly. A high minimum self-bond protects against cheaply buying two thirds of voting power, but the chain does not yet enforce that floor in consensus; it is a proposed figure and an operating discipline, which you can confirm by reading the genesis file. The same page also leaves open, in plain view, who pays the shared network’s signers.
Try it, then tell us
Point your tools at both branches, ask each chain the two questions, and post what you find here in Developers: measurements, bugs, and disagreement with evidence.
Previous: Part 5 · Two foundations: why 9Chain is being built twice | Next: Part 7 · LOVE9, explained
This forum is a community space. An announcement counts only when it also appears on the official channels listed at docs.9chain.org.